Your accessibility overlay is a liability, not a fix — the 2026 WCAG, ADA, and EAA evidence

The one-line-script promise is the 2026 accessibility trap

A whole market now sells accessibility as a snippet you paste into your header: an overlay widget that promises to make any site compliant in minutes. The pitch is seductive because it is cheap, fast, and sounds like insurance. The 2026 data says it is the opposite — it is a liability marker that plaintiff firms and regulators actively look for.

This is not a moral argument dressed up as compliance. It is a measurable one. Below is what actually fails on the web, what the lawsuits cost, what the EU now enforces, and the honest counter-evidence on where the real risk sits — so you can spend on remediation instead of a widget that gets you sued anyway.

What actually fails on the web (WebAIM, February 2026)

The WebAIM Million study of the top one million homepages found that 95.9% had detected WCAG failures in February 2026, averaging 56.1 errors per page — up 10.1% from 51 the prior year. The web is not slowly getting more accessible; on this measure it got measurably worse.

The most useful finding is concentration: six error types have accounted for roughly 96% of all detected errors for seven consecutive years. You do not need a thousand-point audit to move the needle. You need to fix six things, in order, in the underlying code.

• Low-contrast text — 83.9% of homepages (averaging 34 instances per page).
• Missing alternative text — 53.1%.
• Missing form labels — 51%.
• Empty links — 46.3%.
• Empty buttons — 30.6%.
• Missing document language — 13.5%.

The US litigation reality, by the numbers

Seyfarth Shaw's March 2026 analysis counted 3,117 federal website-accessibility cases filed in 2025, up 27% from 2,452 the year before and representing 36% of all ADA Title III filings. They clustered in New York (1,021), Florida (961), and Illinois (585). Roughly 70% of ADA web lawsuits target e-commerce sites, and 35.8% of the top 500 e-commerce retailers have received at least one.

The cost is layered, not a single number. Typical settlements run $5,000 to $75,000 before legal fees; demand letters often open around $5,000; DOJ civil penalties can reach $75,000 for a first violation and $150,000 for subsequent ones — and none of that includes the remediation itself, which is the part you pay for either way.

• 3,117 federal website-accessibility cases in 2025, up 27% (Seyfarth Shaw).
• ~70% of ADA web lawsuits target e-commerce.
• Settlements typically $5,000–$75,000 before fees; civil penalties up to $150,000.

Why overlays do not hold up — and the FTC agrees

The single most instructive statistic in the litigation data: in the first half of 2025, 22.6% of ADA lawsuits (456 cases) targeted websites that already had an accessibility widget installed. Other trackers put the share of sued companies with an overlay at more than a quarter. The widget did not prevent the lawsuit — in several analyses it correlated with being targeted, because plaintiff testing goes straight past the overlay to the unlabeled forms, broken keyboard flows, and missing alt text in the underlying HTML.

Regulators have moved too. The US FTC brought an enforcement action against overlay vendor accessiBe over claims that its widget could make a website WCAG-compliant, resulting in a $1,000,000 settlement and an order barring the claim that an automated product makes any site compliant without evidence. No US court has accepted an overlay as a valid compliance defense.

• 22.6% of H1-2025 ADA lawsuits (456 cases) hit sites that already had a widget.
• FTC: $1,000,000 accessiBe settlement; order bars unsubstantiated compliance claims.
• No US court has accepted an overlay as a compliance defense.

The EU added a second front: the European Accessibility Act

Since June 28, 2025, the European Accessibility Act (Directive (EU) 2019/882) has applied to covered services sold to EU consumers — e-commerce, banking, telecommunications, and digital media among them. The technical baseline is EN 301 549, which incorporates WCAG 2.1 Level AA. Penalties are set nationally and vary widely: up to €100,000 in Germany (plus a private cease-and-desist mechanism that has already produced more enforcement actions than any other member state), €1,000,000 in Spain (with operational bans up to two years), €900,000 or 10% of turnover in the Netherlands, and turnover-based exposure reaching roughly €1,260,000 in Hungary. Ireland is the only member state with criminal sanctions, including up to 18 months imprisonment.

One exemption matters and is widely misread: microenterprises with fewer than 10 employees AND annual turnover below €2 million are exempt from the service requirements. The AND is the trap — a company with eight employees and €5 million in revenue does not qualify. The exemption also covers services only, not products.

• EAA applies since June 28, 2025; baseline is EN 301 549 (WCAG 2.1 AA).
• Fines vary by state: Germany €100k, Spain €1M, Netherlands €900k or 10% turnover.
• Microenterprise exemption needs <10 employees AND <€2M turnover — not either.

The honest counter-evidence: do not panic-buy

The hype says fines are raining down. The verified record is more measured. As of June 2026, no confirmed fine issued specifically under a national EAA implementation could be verified from reliable public sources; early enforcement has run through complaints, notices, injunctions, market surveillance, and corrective-action deadlines rather than penalty checks. Most authorities issue a remediation order before any financial penalty, and several countries allow a cure period (Italy gives 90 days).

On the US side, the DOJ has not adopted WCAG as the binding Title III standard for private businesses. A February 2026 Statement of Interest in Alcazar v. Fashion Nova explicitly declined to endorse WCAG as the necessary Title III standard, and Title III still runs on the general effective-communication obligation plus case law rather than a named web rule. The real, present risk is a high-volume demand-letter environment and e-commerce exposure — not a guaranteed regulatory fine. Spend accordingly: remediate the code, document it, and skip the widget.

• No confirmed EAA-specific fine verifiable as of June 2026; enforcement is corrective-first.
• DOJ has not adopted WCAG as the Title III standard (Alcazar v. Fashion Nova, Feb 2026).
• Present risk is demand letters and e-commerce exposure, not a certain fine.

The compliance playbook that actually holds up

Pin a version: WCAG 2.1 AA is the legal floor in most contexts; WCAG 2.2 AA is the better target and current best practice, and building to it avoids rework as the baseline shifts. Then fix the six WebAIM errors first — contrast, alt text, form labels, empty links, empty buttons, document language — because they account for the overwhelming majority of detected failures and are the ones plaintiff testing catches.

Test the way the failures actually surface. Automated tools catch roughly 30% of WCAG violations; the remaining ~70% require human judgment — real keyboard-only navigation and real screen-reader passes through every form, filter, cart, and checkout. Publish an accessibility statement (France levies a separate €25,000-per-year penalty for not having one), name an accountable owner on the team, involve people with disabilities after — not instead of — your automated and manual checks, and document everything. There is no such thing as ADA certification; a scanner score is not a badge.

• Pin WCAG 2.2 AA as the target; 2.1 AA is the floor.
• Fix the six WebAIM errors before anything exotic.
• Automated tools catch ~30%; the rest needs keyboard + screen-reader testing.
• Publish an accessibility statement; name an accountable owner; document remediation.

Accessibility is also a market, not only a legal duty

The compliance framing undersells the commercial case. The W3C business case counts more than one billion people — about 15% of the world population — living with a recognized disability, controlling over $6 trillion in spending power, with an extended market of 2.3 billion people and roughly $6.9 trillion in incremental disposable income. The CDC puts it at 1 in 4 US adults. A UK study cited by Section508.gov finds 71% of web users with a disability will leave an inaccessible site — they do not file a complaint, they go to a competitor.

Accessible content also performs. In the W3C's worked example of adding transcripts, search traffic rose 6.86% and unique visitors rose 4.18%. The teams that treat accessibility as a design-system constraint — semantic tokens, contrast ratios enforced in CI, keyboard-complete components by default — get the legal risk reduction and the larger addressable market in the same pass. That is the difference between buying a widget and building a product.